' Skip To Content
Influenza 1918 | Article

Interview: Jeffrey Taubenberger

Dr. Jeffery Taubenberger is a virologist at the National Institute of Allergy and Infectious Diseases. In March 1997, a team of pathologists led by Taubenberger published a report in the journal Science concerning the discovery of genetic material from the influenza of 1918.


This interview was conducted in January of 1998. At the time of this interview Dr. Taubenberger was a pathologist at the Armed Forces Institute of Pathology in Washington, D.C., and he headed the institute’s division of molecular pathology.

Mission of Armed Forces Institute of Pathology
The Armed Forces Institute of Pathology, or AFIP, has been around for about 130 years. It came into being during the Civil War, was actually created from an executive order of Abraham Lincoln to the Army Surgeon General to study diseases in the battlefield. And even during Civil War, infectious diseases were a major problem for the military and more people died of infectious diseases than they did of bullets. That’s probably true of all wars. Pathology as a subspecialty of medicine was sort of coming into its own at that time in the late 1800s, and people were beginning to examine tissues under the microscope — tumors and infectious agents and things. So the Institute sort of was created to collect and study disease specimens.

This institute has continued to collect interesting medical specimens, both surgical pathology and autopsy pathology specimens for this period of time. So currently it houses, I don’t know how many millions of specimens, about three million specimens covering basically all ranges of human and actually veterinary pathology.

Genetic Analysis
I’m in the Molecular Pathology Group, and what we do is a variety of basic research and clinical tasks that relate to molecular pathology. What that means is looking at disease through molecular genetic changes. Traditional pathologists would make diagnoses by looking under the microscope. We try to make, or supplement making diagnoses by looking at gene changes.

So our task, in general, for our clinical duties is to actually analyze genetic changes in patient samples that are submitted. Most of these samples are submitted the way pathology samples are traditionally processed. They’re fixed in chemicals like formaldehyde and then imbedded in wax. So it becomes very difficult to isolate and analyze DNA or RNA from this kind of tissue, since it’s very degraded. So we’ve put a lot of effort into optimizing methods to isolate nucleic acids from this kind of tissue. And particularly, the ability to get RNA, which is much more fragile than DNA, out of this kind of tissue has been a big task of the lab. It was optimizing a technique that would allow us to do that, that was the basis for the idea of why we should do the 1918 influenza project.

New Look at an Old Flu
The 1918 influenza virus caused a pandemic that killed minimally 20 million people and up to 40 or 50 million people in about a year’s time. There are no isolates virus available. Influenza viruses weren’t known to exist in 1918. So there’s really no direct information about the virus. It’s still a big medical mystery. So the question was would the Institute possibly have tissue from victims of the 1918 flu and could we use these new powerful molecular technologies to analyze something about the virus to actually get a first direct look at the virus.

New Case Study
It turns out there were about 70 cases in the Institute’s archives that represented autopsies of 1918 flu victims that contained tissues that were formalin fixed and paraffin imbedded as well as accompanying glass slides and clinical records. So we started the project by pulling half of them randomly.

The first thing I did was to examine the cases under the microscope and try to learn something about the pathology of the 1918 victims. And looking at the clinical records and the slides from these cases as well as what’s known about the pandemic in general, people died in different ways. Some people died very quickly after coming down with flu symptoms in 1918, which was something that was really very unusual about the 1918 epidemic. Some people would die within two or three days after the onset of symptoms, and very unusual pathology in which their lungs were just filled with blood, pulmonary hemorrhage where they would just fill with fluid, that is pulmonary edema, and they would basically drown, and with very little inflammation. And this kind of pathology was really unique.

The other really odd thing was that people who died with this kind of pathology tended to be young, healthy people. Influenza viruses normally only kill the elderly or very young, people who basically don’t have normal immune systems. So that was a very unique feature of the 1918 flu.

The reason that this was important is that influenza viruses are known to replicate extremely quickly. That is, they make copies of themselves in the cells that line the lung and then after about five days, they leave that host after having replicated and then infect other people. So after about a week’s time, even under modern circumstances, it becomes impossible to recover the virus. So we wanted to select four cases which the clinical course from onset of symptoms to death was under a week.

One of those cases still contained genetic fragments of the 1918 virus, and that was a case that had really good histologic evidence, that is under the microscope, of primary influenza pneumonia.

Reassembling a Mosaic
Currently we have about 15 percent of the virus decoded. The influenza virus is broken up into eight different RNA segments. It’s not one solid piece of RNA, and they vary from about 1,000 to 2,500 bases in length. The pieces that we can recover from the cases are only about 150 or 160 bases in length. So what we have to do is basically put together these pieces like reassembling a mosaic, one piece at a time, and sort of march along reassembling.

For the initial look, we looked at five different genes, just small segments of five different genes just to get a feel for what the virus in general looked like. But since then, since the publication of the paper, we’ve been concentrating on the hemagglutinin gene, which is the major surface protein of the virus, which is important in how the cell infects humans and is the major immune response to virus. We have actually completely finished reassembling the sequence of that gene, which is about 1800 bases long. We’re doing final analysis on that now and we should submit paper for publication on that analysis in a couple of weeks.

But we really think the material is representative and ultimately we might be able to decode the entire virus. But it’s going to take years of effort because the pieces are so small that to reassemble them is an incredibly tedious and time consuming task.

Risk of Infection?
In performing this research and analysis of the 1918 virus, we don’t have any infectious disease risk of escape of live 1918 virus at all. Indeed, the sample has been fixed in chemical formaldehyde and processed for pathology which would destroy all viruses, including influenza virus. So there’s no way that we could recover live virus. The genes of the virus, as I said, are chewed up into little tiny pieces. So it’s absolutely clear that there’s no live virus, there’s no infectious risk at all.

Flu Clues
The one thing we can say is that this is definitively an influenza virus, a type A influenza virus of the subtype H1N1. Influenza viruses come in three different types — A, B and C — of which type A is the major pathogen for humans. But it also infects multiple other animals, including pigs and birds, chickens, ducks, for example. And the influenza viruses are classified based on the different hemagglutinin and neuraminidase proteins on the surface of the viruses. And there are 14 hemagglutinin type subtypes and 9 neuraminidase subtypes. So this particular virus is so-called H1N1. That was predicted based on the antibodies that people had in their blood who were alive in 1918 and infected with the virus but survived.

The other thing that we can say, which is very interesting, was an observation that was made back in the 1930s using this kind of analysis, which was that swine or pig influenza viruses and human influenza viruses were both discovered in the 1930s. Scientists doing analyses back then of antibodies in the serum of 1918 survivors suggested that the antibodies floating in the bloodstream of 1918 survivors was a better match to the influenza viruses in pigs in the 1930s than it was to the influenza viruses that were in circulation in humans in the 1930s. So that led people to think that this was a so-called swine flu.

Our work also seems to confirm that, that the hemagglutinin sequence looks more like the 1930’s swine hemagglutinin sequences than it does the human hemagglutinin sequences of the 30’s. So there seems to be evidence that there’s some relationship to the classic swine flu.

One thing that is an ultimate goal of the project is to try to figure out how it is that this virus got into humans and started a pandemic. Influenza viruses can infect multiple animals, and it’s thought that the natural reservoir for influenza viruses are wild waterfowl, like ducks. And in general, it’s not thought that avian influenza viruses like those that infect ducks can directly infect humans. In general, it’s thought that influenza viruses have to be adapted to life in mammals before they can infect humans, that the viruses have to mutate so that they change the way in which they bind to cells and replicate so that they can successfully grow in human cells, and that pigs may serve as an intermediary, as a mixing vessel between avian-like viruses and human-like viruses.

One thing that this 1918 virus suggests is that all the gene segments we’ve looked at so far look like they fall into the mammalian class of viruses, that swine/human class, not the avian class. So our analysis suggests that the virus did not come directly from birds to humans in 1918, but went through some kind of mammalian intermediary and pigs are a possibility. Although it’s really rather curious as to actually whether pigs infected humans with the 1918 virus or humans infected pigs in 1918. It was clear that there were outbreaks in both humans and pigs almost simultaneously. So it will be really interesting to try to unravel that, as to how the virus actually got into humans.

A Recombinant Virus
There is some evidence to suggest that humans had been infected with similar H1N1 virus as early as about 1908, by antibody evidence. So it’s possible is that what we’re dealing with is a recombinant virus that has some human-like characteristics of a human virus that was maybe not a killer virus but was already a human adapted influenza virus that was circulating in the decade or so before 1918 and that something happened to it, either by recombination with other viruses in pigs or from birds, and that it turned into a killer virus. What we’re trying to do is to pinpoint genetic features of the 1918 virus that might explain its unusual virulence.

Seeds of a Pandemic
Influenza viruses are known to continually circulate in humans and they cause small outbreaks every winter. And with great regularity, new influenza viruses arise in human populations that spread throughout the world very quickly and cause a so-called pandemic, which is a major epidemic. These happen about every 10 to 30 years, if you go back at least to the records of the 1700s, and probably have happened throughout human history. What happens here is that a completely new kind of influenza virus emerges.

Influenza viruses are not stable genetically. They’re really, in a sense, chameleon viruses. They can change their coats. And this is presumably an adaptation of the virus, to evade the host immune response, so that the influenza virus that was circulating last year is not the same as the influenza virus that’s circulating this year. That’s why they have to reformulate the vaccine preparations every single year and people have to be reinoculated, because the virus mutates and changes enough so that your prior immunity would not be sufficient to prevent you from being infected with the virus. So they’re very clever in that sense.

These mutations tend to be sort of slow and steady and accumulate every year. Occasionally, however, a really dramatic change occurs in which an entirely new kind of influenza virus emerges that hasn’t existed before. Usually it’s thought that it’s a recombinant. That is, it’s a mixture of the genes of two different influenza viruses. So that if one animal, say a pig, is infected with a bird influenza virus and a human influenza virus, what could happen is that you could make a mixed virus, a recombinant virus that had some genetic material of the bird virus and some genetic material of the human virus, and it would be so different that no one on earth would have any kind of immunity to it and it would allow it to spread like wildfire through human populations.

So presumably, that’s what happens in pandemics, and there have been three of them in this century. In 1918, in 1957, and 1968. So it’s been 30 years since the last one occurred. So if you want to bet the odds, it’s very likely that another pandemic is really due.

So one of the goals of the project with 1918 is to not only think about what might be the genetic basis for a virus that was so virulent, that was really different in its behavior from other known influenza viruses, and also how it could have emerged, so that we can use this information to look for the emergence of new influenza viruses in pigs or in humans that actually may be the source of a new pandemic.

It’s kind of counterintuitive sometimes, that an agent that — an infectious agent that is very good at killing its host is actually not a very well adapted infectious agent. That is, that bacteria that live on your skin or inside your colon are extremely adapted to living in humans. They don’t cause disease. And so they have a nice relationship. They have set up shop and they live happily on your skin and you don’t do anything about it, and occasionally there’s even a mutually beneficial relationship that occurs. For example, some of the bacteria in your colon make a vitamin, vitamin K that we need.

In the case of bacteria or viruses that are not so well adapted, if they kill their host very quickly, they actually may limit their own ability to spread to other people. That is, if they kill people so quickly that they don’t have time to replicate, they’re kind of hurting their own chances of replication.

In this case, it was something in between. It was a virus that certainly killed a lot of people, but realistically only killed a small percentage, 2 to 5 percent of the people that were infected. So it was clearly a really good virus to infect humans. It spread amazingly well from person to person and spread all over the world. But probably in about a year’s time, from 1918-19, practically every person on earth was at least exposed to this virus. So everyone formed an immune response against the virus in one degree or another. So that would put enormous pressure on the virus to mutate, to change its coat so that it could continue to infect humans.

Possibly what happened was that whatever caused this virus to be so unbelievably virulent changed when the virus mutated and the viruses that continued to circulate both in humans and in pigs after 1918 are probably actually mutated descendants of the virus but not the 1918 virus itself.

Influenza forecasting is very important to the medical community since influenza viruses, even in normal epidemic years, kill thousands of people around the world. So the prediction for new influenza viruses is very important. So the World Health Organization and the Centers for Disease Control here in the United States and other international health agencies have monitoring stations around the world to analyze samples of influenza as they appear.

Historically, it seems that most new influenza viruses emerge in Asia, in the Far East, which is another thing that’s unusual about the 1918 virus because everything we know historically suggested that it actually originated in the United States.

But it’s thought that new influenza viruses emerge by this recombination mechanism, where different viruses are mixed together. And in the Far East, you have sort of cultural lifestyles in which ducks and pigs and humans co-habitate much more closely than they would in the West. So it’s thought that that milieu actually fosters the production of new viruses.

So most of their monitoring actually occurs in China and other parts of the Far East. And then these laboratories isolate and analyze these viruses to try to learn how they differ from the previously recognized viruses. And if an incredibly different virus emerges, then they would have to fear that this might be the emergence of a new pandemic.

Norway Expedition
The Norway expedition was initiated by Dr. Kirsty Duncan in Canada, who read Dr. Alfred Crosby’s book on the history of the 1918 flu and thought that it would be extremely useful to find a source of the 1918 virus so that genetic analysis could be made. She looked at sources in which frozen bodies might possibly still be, Alaska being one of them, and finally found out about seven miners who were buried on the island of Spitzbergen which belongs to Norway, which is inside the Arctic Circle, in which these miners died in the fall of 1918 of influenza. There had been prior expeditions to Alaska in the 1950s to exhume frozen bodies of victims of the 1918 flu with the attempt to culture the virus.

In the 1950s influenza viruses could be cultured and characterized in the laboratory, but molecular genetic analysis like we’re doing now was not possible.

So bodies were exhumed and examined and attempts to culture virus were made, which were unsuccessful. They were able to actually culture out some live bacteria still from the lungs of these victims, but no influenza viruses.

So her goal is to repeat this but to use some of these more modern molecular techniques as well. This project was initiated completely independently of our work and vice versa. Probably initiated about the same time and we now are in contact and know one another and I have participated in meetings about the Spitzbergen material and it’s possible that our laboratory here will analyze tissues for genetic evidence of the 1918 virus should exhumation in the Spitzbergen island occur.

Goals of Norway Team
One of the stated goals of that team is to attempt to culture live virus from these samples. I think it’s extraordinarily unlikely. So unlikely that it’s almost unimaginably unlikely that live virus would be recovered. Influenza viruses are very fragile. And even if the bodies were preserved and frozen as they are in permafrost to retain some RNA of the virus, the genetic material, I think it’s almost inconceivable that live virus would be obtained. It was the case in the 1950s exhumations, when the samples were only half as old as they are now, they were only 38 years old, that no live virus was obtained.

Anyway, they are going to attempt that. Those attempts would be done under extremely tight biosafety measures. That is, there are only a certain few laboratories in the world that have these extreme biosafety containment laboratories. The CDC in Atlanta being one of them. Fort Detrick in Maryland, the Army’s facility, being another.

Outbreak in Hong Kong
The recent occurrences in Hong Kong are really interesting, in which you have a limited number of people who have been exposed to truly avian influenza virus. That is, all the gene segments that have been analyzed so far, to my knowledge, really look avian like. This is not a recombinant human avian virus.

So it really looks like the first time in a natural setting that humans have been infected with an avian influenza virus. It looks, however, like there is no evidence of person-to-person spread of the virus. So somehow you have an avian influenza virus that has been able to gain a foothold and replicate to at least some extent in human beings, but that the virus is not well adapted to life in humans and consequently can’t really spread from human to human. But since influenza viruses mutate so rapidly, that possibility certainly exists. And if this virus was able to be passed from person to person, the risk of it becoming a pandemic virus will be very great because it’s a very different virus than any other viruses that have circulated in humans in this last century. So it would be the same situation like 1918. No one would have immunity to this kind of virus and it really would spread all over the world.

Preparing for a Pandemic
We are in much better shape than we were in 1918. First of all, we know that influenza viruses exist and we can analyze them and watch their emergence and evolution. Secondly, health care has come a long way — the consequences of influenza infection, including drugs that inhibit influenza virus replication directly. Secondary consequences like bacterial pneumonia, we have antibiotics and other supportive therapy that would be used to treat that. We have the influenza vaccines, which are obviously the most important factor of our current armamentarium against influenza viruses. So all of those things could be put into motion.

The problem would be to do it fast enough. Production of a vaccine takes at least six months and it’s having that lead time to get enough vaccine made would be the difficult issue. Another complicating factor with the Hong Kong situation is that this is really a virus that is adapted to life in chickens and is very lethal to chickens, and influenza viruses are actually grown in chicken eggs to make the vaccines. This was a procedure worked out in the 1930s and it really is still the gold standard in which vaccines are made. But this virus actually kills the chicken eggs, and so they wouldn’t be able to make a vaccine directly against this virus using the traditional mechanisms.

Speed Kills
The spread of influenza is remarkably rapid. And looking at pandemics and descriptions of them going back to the 1700s, in which there must have been a lot less movement of people around the world, still spread amazingly rapidly. In 1918, the virus clearly spread within a period of a month or so in the fall all over the world. Certainly the movement of people and transports and troops in response to World War I was probably a factor, but the other pandemics that have occurred in this century, in ’57 and ’68, were not during war time. So these viruses spread just as rapidly all the way around the world.

So I think that you would say that maybe it’s worse now because you have jumbo jets coming from Hong Kong to major cities all over the world every day, that that would probably help the spread. I think that even without modern transportation viruses spread, these viruses spread quickly enough. So it would be a real worry. I mean, the virus really, if it got a foothold in humans and could be passed from human to human, it really could spread around the world within a period of months.

Emerging Infectious Diseases
Certainly this idea of so-called emerging infectious diseases is a big concern in the medical community. There have been, I guess really with the advent of HIV probably in the 80s, this idea that new and uncharacterized infectious agents really were emerging. I think that, in general, that there was this perception that because of medical advances in the last hundred years that infectious diseases were really a thing of the past, that it wasn’t very interesting. We had discovered all the viruses and bacteria, we have drugs that can fight them, antibiotics, and now we should concentrate on cancer and heart disease. Well, it turns out that infectious agents and nature are still are a lot cleverer than we thought, and clearly we have not won this battle about infectious agents.

I think that the huge population growth in the world and the expansion of humans into remoter areas of the world and the spread of humans from one place to another has clearly increased the ability of infectious agents to gain a hold in human populations. So things like Ebola virus and Hanta virus that are incredibly nasty viruses that really kill people, but in really isolated places, still exist. But we shouldn’t forget things like influenza viruses, which sound very boring and commonplace, still have the potential to cause great medical harm.

So I just don’t think we should forget the lessons of microbiology, that infectious agents are constantly trying to figure out ways to infect people and that we really can’t consider this battle won.

New Flu Due?
Just looking at the clinical records of the past hundred years, I think it’s extraordinarily likely that another influenza pandemic will occur. As I’ve said, they seem to occur with great regularity every 10 to 30 years, and so it has been 30 years since the 1968 pandemic. So the odds are very great, practically a hundred percent, that another pandemic will occur.

Whether the current Hong Kong virus is a virus that has pandemic potential is not yet known, but it’s clearly being looked at very, very carefully. So if it really does gain that potential to spread from human to human, I think this will be learned very, very quickly.

Support Provided by: Learn More